01:07 < zoq> robb8: https://github.com/mlpack/mlpack/blob/master/src/mlpack/tests/feedforward_network_test.cpp is one example

01:08 < zoq> robb8: https://github.com/mlpack/models is another

04:15 < jenkins-mlpack2> Project docker mlpack nightly build build #278: FAILURE in 1 min 34 sec: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/278/

08:00 < jenkins-mlpack2> Yippee, build fixed!

08:00 < jenkins-mlpack2> Project docker mlpack nightly build build #279: FIXED in 3 hr 37 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/279/

12:34 < travis-ci> shitian007/mlpack#2 (k_means_parallel - 4c05287 : Shi Tian): The build has errored.

12:34 < travis-ci> Change view : https://github.com/shitian007/mlpack/compare/3f918cf82e4a^...4c05287e39dd

12:34 < travis-ci> Build details : https://travis-ci.org/shitian007/mlpack/builds/513434808

14:50 < sreenik> Hi, this might be a weird question here. I use Visual studio code for C++ programming. However, the memory it consumes increases slowly over time and as it reaches the 7-8 hour mark the system consumes around 7gb of RAM (around 2.5gb is the system consumption when I start VSCode). Could this be due to a memory leak in the programs that I run or is visual studio code at fault? I use linux

14:56 < zoq> sreenik: Could be a memory leak, perhaps you can track this down to a program you run.

14:58 < sreenik> zoq: Thanks. Unfortunately this happens with the cli parser I am developing. Ah, more digging to do... :)

15:00 < zoq> sreenik: Running valgrind might be helpful here.

15:02 < sreenik> Thanks I'll try that :)

16:48 < whoKnows> once I did some changes in the source code do I need to click "build solution" under build tab in MVS or any other option beacuse by doing build solution every time it takes a very long time to build the project even for a small change in source code

16:49 < whoKnows> Can't I build project only for the specific changes I did keeping everything intact

17:10 < sreenik> whoKnows: I don't know if this will solve your issue but you can do your changes and testing in a different workspace and if everything seems to work fine, just do similar changes in your original project and build it.

17:16 < sreenik> You can actually build individual components but I am not sure if this is exactly what you want to do. Instructions are here https://webcache.googleusercontent.com/search?q=cache:q3PxVxTLoKcJ:https://www.mlpack.org/docs/mlpack-1.0.9/doxygen/build.html+&cd=2&hl=en&ct=clnk&gl=in (this is a cached copy of the old site, I have currently forgotten the name of the new site)

17:22 < whoKnows> I think this may be helpful for the next time but I just wanted to know that everyone suffers such long time for building entire project or I am doing something wrong because in my system I took an average time of around 45 minutes which is very huge

17:42 < sreenik> zoq: There are many optimizers in ensmallen that I have never heard of. Are all of them applicable for ANNs or do many of them (only) have other general purpose uses?

18:32 < zoq> sreenik: Every optimizer that can handle differentiable separable functions, can be used for ann's: https://ensmallen.org/docs.html#differentiable-separable-functions

18:47 < sreenik> zoq: ahh I see that I have implemented a couple of the non-differentiable functions also. Thanks, your reply was just in time :)

18:50 < zoq> sreenik: Great :)

22:05 < robb_> Hey, so I'm not entirely sure what the 3rd dimension for recurrent neural networks is.

22:05 < robb_> I've read the documentation and it looks like each slice is a data point

22:06 < robb_> oops, I mean time step

22:06 < robb_> and each column is a data point. so would each row be a feature?

22:50 < saksham189> robb_: you are correct about each slice which refers to a time step. each slice has multiple data points (columns) for a given batch which leads to rows

22:51 < saksham189> hope i am clear

23:08 < robb_> saksham189: thank you for the clarification. so each row is a batch?

23:11 < saksham189> a row corresponds to a single dimension of multiple data points for certain time step

23:15 < robb_> ok thank you. i just wish there were better RNN examples . I had to pull the sine wave one from the boost tests lol

23:15 < robb_> still trying to figure it all out